TW201539551A - Monitoring method and apparatus for control of excimer laser annealing - Google Patents

Abstract

A method is disclosed evaluating a silicon layer crystallized by irradiation with pulses form an excimer-laser. The crystallization produces periodic features on the crystalized layer dependent on the number of and energy density ED in the pulses to which the layer has been exposed. An area of the layer is illuminated with light. A microscope image of the illuminated area is made from light diffracted from the illuminated are by the periodic features. The microscope image includes corresponding periodic features. The ED is determined from a measure of the contrast of the periodic features in the microscope image.

Description

Monitoring method and device for control of excimer laser annealing Related application

The priority of the US Patent Application No. 14/195,656, filed on March 3, 2014, is hereby incorporated by reference.

Technical field of the invention

The present invention relates generally to the melting and recrystallization of a thin tantalum (Si) layer by pulsed laser irradiation. This method is particularly relevant to the evaluation of such recrystallized layers.

Description of the background art

Rhodium recrystallization is a step that is often used in the manufacture of thin film transistor (TFT) active matrix LCDs, and organic light emitting diode (AMOLED) displays. The crystallization enthalpy forms a semiconductor substrate in which the electronic circuitry of the display is formed by conventional lithography. In general, recrystallization is performed using a pulsed laser which is formed into a long line and has a uniform intensity profile along the length direction (long axis) and also has a profile in the width direction (short axis). Uniform or "top hat" strength profile. In this process, a thin layer of amorphous germanium on a glass substrate is repeatedly melted by the pulse of the laser radiation, and the substrate (and the layer of germanium thereon) is The source translation is transmitted relative to one of the laser radiation pulses. By repetitive pulses at a certain optimum energy density (OED), melting re-solidification (recrystallization) occurs until a desired crystal microstructure is obtained in the film.

Optical elements are used to form the laser pulses into a line of radiation, and recrystallization occurs in a strip having the width of the line of radiation. Each attempt will be used to maintain the intensity of the radiation pulses consistently along the line. This is necessary to maintain the uniformity of the crystal microstructure along the strip. One of the preferred sources of light pulses is a quasi-molecular laser that transmits a pulse having a wavelength in the ultraviolet region of the electromagnetic spectrum. The above recrystallization method using excimer laser pulses is generally called excimer laser annealing (ELA). This method is very sensitive, and the error tolerance of the OED will be a small percentage, or even as small as ±0.5%.

There are two modes of ELA. In one mode, the translational speed of a panel relative to the laser beam is sufficiently slow, such that the "top hat portion" of the beam width overlaps by up to 95% from one pulse to the next, so any minimum The area will accept a total of approximately 20 pulses. In another mode called progressive ELA or AELA, the translational speed is much faster, and the "straight lines" of the illuminations have the least overlap when passing through one panel at a time, and may even leave no Crystallized space. Multiple passes are made to cause the entire panel to be illuminated by a total number of pulses, which can be made in an equivalent material in an ELA process.

Regardless of which ELA mode is used, the evaluation of the crystallization mold on the panel in a production line is immediately performed offline by visual inspection. This check is completely subjective and relies on highly trained and experienced Investigators, by their experience, are able to correlate the observed panel characteristics with very small changes in energy density in the crystallized beam, for example less than 1%. In a production environment, if there is a change in the process energy density, the visual analysis and establishment process typically takes about one to one and a half hours from the time the crystallization is performed, and the production capacity of the acceptable panel is There is a corresponding negative impact.

Therefore, there is a need for an objective method of evaluating the ELA process. Preferably, the method should be practiced on at least one production line. More preferably, the method should be capable of being used for real-time evaluation in a feedback loop and automatically adjusting the process energy density in response to the information provided by the evaluation.

Summary of the invention

The present invention relates to a method and apparatus for evaluating an at least partially crystallized semiconductor layer that is exposed to a plurality of laser radiation pulses having an energy density on the layer, the crystallization being Forming, on a layer, a periodic surface feature of a first group in a first direction and a periodic surface feature of a second group in a second direction perpendicular to the first direction, the first and the The form of the periodic characteristics of the two groups depends on the energy density of the laser radiation pulse to which the semiconductor layer is exposed.

In one aspect of the invention, a method for evaluating a semiconductor includes: illuminating a region of the crystallized semiconductor layer, and periodically entanglement of the first and second groups from the illumination region A microscope image of one of the illumination regions is recorded in the shot light. The recorded image contains periodic features corresponding to the first and second groups of illumination regions of the layer, respectively Horizontal and vertical periodic image feature groups. The energy density is determined by one of the comparisons of at least one of the horizontal and vertical periodic image feature groups.

20‧‧‧Evaluation device

21‧‧‧Diffraction measuring device

22‧‧‧ Crystallized layer

24‧‧‧glass panel

26‧‧‧Microscope

28,80‧‧‧Light source

29,65,82‧‧·beam

30‧‧‧Poly partition

32‧‧‧Optical components

34‧‧‧Reflected beam part

36R, 84‧‧‧Diffraction

36T‧‧‧Diffraction beam part

38‧‧‧stops

39,40‧‧‧ Filter elements

50‧‧‧Detection unit

52‧‧‧Light detection components

54‧‧‧Electronic processor

60,60A‧‧‧ Annealing device

62‧‧‧ translation stage

64‧‧‧Excimer laser

66‧‧‧Variable attenuator

68‧‧·beamforming light parts

69‧‧‧Forming beam

70‧‧‧Rotating mirror

72‧‧‧Projected light pieces

82R‧‧·reflected beam

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in FIG The principles of the invention are illustrated.

Figure 1 is a diagram schematically showing fast Fourier transforms (FFTs) of scanning laser microscope images for ELA crystallized germanium layers. The highest amplitude measured in the rolling and transverse directions is a function of pulse energy density. .

2 is a graph schematically showing FFTs of a scanning laser microscope image for an A-ELA crystallized layer, the highest amplitude measured in the rolling and transverse directions being a function of pulse energy density.

3 is a graph schematically showing fast Fourier transforms (FFTs) of scanning laser microscope images for A-ELA crystallized germanium layers, the highest amplitude measured in the rolling and lateral directions being one of the number of pulses function.

Fig. 4 is a polarized light microscope image of a region of an ELA crystallized ruthenium layer showing ridges formed transversely to and parallel to the rolling direction (RD) of the layer when crystallized.

Figure 5 is a conoscopic microscopy image of a region similar to one of the crystallization layers of Figure 4 showing the horizontal and vertical bands of light from the diffracted light from the transverse and rolling ridges, respectively.

Figure 6 is a diagram schematically showing crystallization from ELA The amplitude measured by the diffracted light of the transverse and rolling ridges of the layer is a function of the pulse energy density.

Figure 7 is a graph schematically showing one of the number of pulses measured for the ED values of 410, 415 and 420 mJ/cm ² of diffracted light from the transverse and rolling ridges of the A-ELA crystallized layer. function.

Figures 8 and 8A schematically illustrate a preferred embodiment of a device for separately measuring the amplitude of diffracted light from the lateral and rolling ridges of an ELA crystallized layer in accordance with one aspect of the present invention.

Figure 9 is a schematic illustration of a preferred embodiment of an ELA device in accordance with the present invention, comprising the apparatus of Figure 8 and having a variable attenuator for responding to diffracted light from lateral and rolling ridges of the ELA crystallization layer The measured amplitude adjusts the pulse energy density on a layer.

Figure 10 is a schematic illustration of another preferred embodiment of an ELA device in accordance with the present invention, similar to the device of Figure 9, but wherein the device of Figure 8 is used in accordance with the present invention to separately measure crystals from the ELA. Another preferred embodiment of the device for laterally and laterally swaying the amplitude of the diffracted light of the ridges is replaced.

Figure 11A is a reproduction of a transmission microscope image recorded in a diffracted light from a region of a germanium layer crystallized below an optimum energy density (OED), the image containing horizontal and vertical periodic characteristics corresponding to The scrolling direction and the lateral characteristics in this layer.

Figure 11B is a reproduction of a transmission microscope image similar to the image of Figure 11A, but as a layer of crystallization that is crystallized above the optimum energy density.

Figure 12 is a graph showing the reproduction of a transmission microscope image similar to that of Figure 11B, and the optical amplitude as a function of horizontal and vertical distance from which the contrast of horizontal and vertical characteristics can be measured.

Figure 13 is a diagram schematically showing the contrast measured in a horizontal feature of a microscope image similar to the image of Figure 12, as recorded by the graph of Figure 2, by the layer characteristics of the two groups. The amplitude of the light is a function of the energy density and the calculated slope of one of the measured amplitudes is a function of the energy density.

Figure 14 is a diagram similar to the graph of Figure 13, but wherein the microscope image is from a layer of ruthenium crystallized by a laser crystallization process comprising a micro-smoothing technique.

Detailed description of the invention

ELA treatment of the tanning film results in the formation of surface roughness protrusions which are formed by the expansion of the crucible upon curing. These protrusions are particularly formed between three or more cured frontal collisions when grown laterally. These protrusions are usually not randomly located. Rather, they are aligned due to the process of corrugation, collectively referred to in the literature as laser-induced periodic surface structures (LIPSS). Therefore, the corrugations are composed of a series of well-arranged protrusions. The corrugation formation is only seen in an energy density window (range) where the film is partially melted. Typically, the ripple periodicity is equivalent to the wavelength of the incident light, for example, with a XeCl excimer laser of about 290-340 nm. Because of these small dimensions, the ripples are not, or extremely difficult to resolve, using conventional optical microscopy techniques.

Typically seen in bright field microscopy, the surface of the ELA treated film consists of an elongated darker colored area interspersed with brighter areas. The close inspection of the darker regions shows that they are composed of (regular) regions with more intense fluctuations of higher protrusions, with regions that are more irregular and/or lower protrusions therebetween. These more regular regions are referred to herein as ridges, and the regions in between are referred to as valleys. It has been found that the formation of the ridges appears to be associated with the formation of corrugations, and the typical orientation of the ridges is in a direction perpendicular to the direction of the corrugations. The method and apparatus of the present invention relies on measuring light diffraction from a ridge in a thin film (layer) formed by the ELA process. The method provides an indirect measurement of the degree of fluctuation that can be used to monitor or control the ELA process in real time. Moreover, one method is described more directly with respect to the corrugations themselves, although some microscopic techniques are used to be slower than the more conventional optical microscopy techniques used to measure diffraction from ridges.

Corrugations are usually not formed in only one direction. The corrugations are mainly formed in a direction parallel to the scanning direction, and also in a direction perpendicular to the scanning direction (the linear direction). The corrugations are periodic and are described herein in the direction of their periodicity, using terms commonly used in metallurgy, where the rolling direction (RD) will coincide with the scanning direction, and The landscape (TD) will match the direction of the line. The edge is that since the corrugations oriented in the scanning direction are periodic in the transverse direction, they are referred to as TD corrugations. Similarly, the corrugations oriented in the direction of the line are periodic in the rolling direction and are therefore referred to as RD corrugations.

According to the LIPSS theory, the TD ripple has a spacing roughly equal to the The wavelength of light, while the RD corrugations are separated by approximately λ/(1±sin θ), and typically predominantly the λ/(1-sin θ) gap, where θ is the angle of incidence of the laser radiation on the layer, Typically in ELA is about 5 degrees or more. Corrugation formation is helpful in obtaining a uniform polysilicon film because the grain structure tends to follow the surface periodicity. When corrugations occur, ideally, a very regular film consisting mainly of rectangular grains having a size of about λ x λ / (1 - sin θ) will be formed. At lower energy densities (ED), the grains will be smaller, while at higher ED, the grains will be larger. When grains larger than the size of the corrugated field are grown, referred to herein as super lateral growth (SLG), surface reflow will result in a reduction in the height of the protrusion and a gradual loss of order in the film.

In a first experiment to determine the numerical relationship between the surface periodicity caused by the corrugations and the ED of the laser pulse, a laser scanning film laser scanning microscope (LSM) image is used for the RD and TD. The transformations made in the direction are analyzed by Fast Fourier Transform (FFT). One of the peaks in the FFT indicates the presence of a periodic surface, and the position of the peak corresponds to the periodic direction of the surface. A TD transform that provides sharp peaks at approximately 1/λ represents a strong TD periodicity. The RD transform will show less sharp peaks at approximately (1-sin θ) / λ and have a lower amplitude than the TD transforms, ie, the less significant RD ripple has approximately (1-sin θ) / The interval of λ.

Figure 1 is a diagram schematically showing a total of 25 overlapping pulses in an ELA process, the amplitudes of the corresponding RD and TD transform peaks being measured in millijoules per square centimeter (mJ/cm ² ) in each pulse. A function of energy density (ED). It can be seen that the RD periodically shows that it is maximal at an ED that is slightly higher than the maximum TD periodicity. Here, an OED of about 420 mJ/cm ² is shown to periodically decrease (relatively) with the higher ED in the RD and TD directions. It should be understood that the ED system defined herein is determined using a common method in the industry, including measuring the energy in the beam and dividing it by the top cap width of the beam, regardless of the sides of the top cap. gradient.

Figure 2 is a graph similar to Figure 1, but crystallized by a 25 pulse A-ELA process. Among them, the RD ripples show a stronger periodicity than the ELA, and their peak periodicity is better defined than in the case of the ELA process.

Figure 3 is a graph schematically showing one of the RD and TD peak amplitudes as a function of the number of pulses of ED at 420 mJ/cm ² which is slightly smaller than the empirically determined OED. It can be seen that the periodicity in the TD direction is steadily increased to a number of pulses of approximately 22. In this RD direction, there is very little periodic growth until approximately 15 pulses have been transmitted.

Figure 4 is a polarized microscope image of a reflected light. The ridges oriented in the transverse direction (which are associated with the corrugations associated with the RD direction, or in other words, according to the definition of periodicity, i.e., the "TD ripples") can be clearly seen. The ridges oriented in the direction of the scroll (and associated with the "RD ripple") are less noticeable but still visible, as would be expected from the FFT analysis described above.

Unlike corrugations, the ridges are not strictly periodic. However, the ridges have a feature spacing which may typically range between about 1.5 [mu]m and about 3.0 [mu]m, or about a magnitude greater than the spacing of the corrugations. According to the specific terminology of the corrugations, the ridges are referred to in the periodic direction, i.e., the RD ridges are oriented in the transverse direction, and the TD ridges are oriented in the rolling direction.

The FFT analysis, in and of itself, clearly provides a means of evaluating a crystallization layer. However, the steps required to generate the above information are generally slow, and the use of such analysis in near real-time online monitoring or evaluation of a layer crystallized with ELA or A-ELA is discouraged. Therefore, it was decided to study the possibility of analyzing the diffraction phenomenon associated with the vertically oriented ridge groups associated with the RD and TD ripples, without attempting to directly measure the corrugations themselves.

Figure 5 is a conoscopic microscope image of a layer as shown in Figure 4. This is taken using a commercially available microscope that is removed to allow an image of the back focus plane of the objective to be recorded. In this example, the image is recorded as a simple mobile phone camera. The microscope is used in a transmitted light configuration. A first polarizer is disposed in the illumination path in front of the sample, and a second polarizer (analyzer) is disposed after the sample, and a polarization direction is 90 for the first polarizer. degree.

The center of the cone image corresponds to the optical axis of the microscope system, and the distance from the optical axis (center point) corresponds to the angle at which the light travels. Thus, the cone image provides information in the direction of the light in the microscope.

A polycondensation spacer is positioned close to a minimum aperture to limit the angular distribution of incident light on the sample and thereby confining the image of the aperture to the center of the cone image. The remainder of the image is formed by light diffracted by the crystallization of the TD and RD ridge groups. The polarizer and analyzer act together to minimize the brightness of the center point relative to the rest of the image. Rotating the two polarizers at 90 degrees will form a pair of intersecting matting strips in the cone image, referred to as matte shadows. By rotating the polarizer and the analyzer relative to the sample, the extinction shadows are Rotate away to minimize the extinction of the bands.

The actual image presented in Figure 5 on a gray scale is a color image. The horizontal band is a light blue color and the vertical band is a light green color. The colors of the bands can be fairly uniform, so it is believed to have a high diffraction efficiency at the wavelengths and a lower diffraction efficiency at other wavelengths. The uniformity of the color of the bands is the result of one of the variable spacing of the ridges. There may be some spectra that overlap between the spectra of the horizontal and vertical bands.

The microscope objective is a 20x objective. The edge of the segment with a higher intensity gradient at one of the center points will give an indication of one of the image pixel sizes. The larger blocks in these dark quadrants are one of JPEG image compression.

In the horizontal direction of one of the figures, a strong band of light is caused by diffraction of the RD ridges (with respect to TD ripples). In the vertical direction of the figure, a weaker band of light is caused by diffraction of the TD ridges (with respect to RD ripple). The transmitted light creates a bright spot in the center of the image.

As can be expected from the graphs of Figures 1 and 2, if the pulse ED falls below the OED, the relative brightness of the brightness of the TD ridge diffraction strip relative to the RD ridge diffraction strip will decrease with the decremented ED. Sharply reduced. When the pulse ED rises above the OED, the relative brightness of the TD ridge diffraction band will remain approximately the same as the brightness of the RD ridge diffraction, but both will be steeply reduced with increasing ED. drop. Therefore, measuring the brightness of the diffractive tapes provides a powerful way to determine if the ED is above or below the OED and how much.

Figure 6 is a graph schematically showing the RD ridge diffraction intensity (solid line) of the 矽 layer region crystallized by 25 overlapping pulses in an ELA process. The TD ridge diffraction intensity (dashed line) is a function of the pulse ED. The strength of the ridges is not directly measured. Rather, one of the measurements of the intensity of the diffractive tape is derived from the observation that the bands have different color and color information that are still present in the regular microscope image.

A commercially available field chart editor is used to determine the average brightness of the blue and green channels of the polarized image as one of the diffraction measurements of the RD ridge and the TD ridge, respectively. One disadvantage of this method is that the image color channel does not provide an optimized filter to see the band brightness, so there is a considerable crosstalk between the two signals. And the signal of the non-diffractive center point is superimposed on the color channels, so that they have a higher noise scale. Even so, the difference clearly shows a trend that when the ratio of the luminance of the green channel to the luminance of the blue channel reaches a maximum value, the OED is found, as in the dotted curve in FIG. Shown.

Alternatively, a cone image recorded in a CMOS array or CCD array, similar to the image of Figure 5, can be processed electronically using appropriate software to collect measurement data from only the diffraction bands. This has the advantage that the measured value will not be sensitive to the actual color in the image and the diffraction efficiency of the diffracted strips, since the spatial information is essentially dry. This actual diffraction efficiency can be a function of film thickness and deposition parameters.

Figure 7 is a graph schematically showing the RD ridge diffraction intensity (solid curve) and the TD ridge diffraction intensity (dashed curve) as the number of pulses and the pulse ED continuously delivered to the same region of a crystallized layer. One of the functions. The trend among them is similar to the chart of Figure 3. The three ED values in each case were 410 mJ/cm ² , 415 mJ/cm ² and 420 mJ/cm ² , i.e., an interval slightly larger than 1% of the ED was selected. It can be seen that after 15 pulses have been deposited, this 1% change in ED will increase the signal amplitude to a change of about 20%. At about 22 pulses, the diffracted signal change is still about 5% or preferably 2% of the ED change. This clearly shows the sensitivity of the method of the invention.

Figure 8 is a schematic illustration of a preferred embodiment 20 of an apparatus for evaluating a layer of crystallization according to the present invention. The evaluated crystallization layer 22 is supported on a glass panel 24. A microscope 26 that is set up for Kohler illumination includes a light or source 28 that transmits a beam 29 of white light. A polycondensation barrier 30 provides control of the numerical aperture of the cone of light for the beam 29.

A partially reflective and partially transmissive optical element 32 (a beam splitter) directs the beam 29 onto the layer 22 such that it is incident perpendicularly to the layer, as shown in FIG. A portion 34 of the beam is reflected by the layer 32 and the portion 36T is diffracted. The suffix T, if used herein, means that the light system is diffracted by the lateral (TD) ridge formed by the crystallization of the layer. Figure 8A shows device 20 in a plane perpendicular to the plane of Figure 8 and showing the above-described rolling direction (RD) ridge diffraction formed when light 36R is crystallized by the layer.

The reflected and diffracted light is transmitted through element 32. This reflected light is blocked by a stop 38. The diffracted light bypasses the stopper 38 and is incident on one of the photodetecting elements 52 in the detecting unit 50. An electronic processor 54 is provided in the detection unit 50 and is arranged to determine the amplitude of the diffracted light received by the detector.

The detecting component 52 can be a pixel operated detector, such as a CCD array or a CMOS array as described above, and one of the diffracted lights is recorded. The cone image (see Figure 7), by which the intensity of the diffracted light can be determined by the processor 54 in spatial analysis. Alternatively, the detection element can be one or more photodiode elements capable of recording the polymerized diffracted light. For the purposes of this example, the selective filter elements 39 and 40 will be provided with a passband selected to correspond to the particular color of the TD and RD diffracted light as previously described. They can be moved into or out of the path of the diffracted light, as indicated by arrow A in FIG.

In either case, another spectral filter (not shown) can be provided to limit the bandwidth of light from source 28 to the diffracted colors. This will reduce the noise generated by the light (not shown) scattered by the layer 22, which can bypass the bypass 38 and mix with the diffracted light.

In Figures 8 and 8A, the light member of microscope 26 includes a collection lens light member for light source 28, an (infinitely calibrated) objective light member, and a tubular lens light member, etc., which are not shown for ease of illumination. In addition, the microscope can be equipped with a Bertrand lens to directly view the cone image and the "eyepiece" (or eyepiece). The form and function of such optical components in a microscope are well known to those skilled in the art, and a detailed description thereof is not required to understand the principles of the present invention.

It can be selected as a reflected light microscope, and a transmitted light microscope can also be used. This microscope design does not have a beam splitter, but a separate condenser lens is required in front of the sample. For best results, the beam stop 38 can be placed in the back focus plane of the objective, or in any conjugate plane after the sample. In the case of reflected light microscopy, the beam stop is preferably placed in a conjugate plane of the back focus plane of the objective. After being tied to the beam splitter, the forcing will also block the incident light.

It should be noted that the diffraction from the ridges can also be observed without a polarizer and/or a beam stop. The diffractive light strip can still be observed after the objective lens and/or the polycondensation lens are removed. Such lenses should therefore be considered a tool that optimizes the measurement of the brightness and selectivity of the region within the film being detected. They are not a critical component of the device described herein.

Figure 9 is a schematic illustration of a preferred embodiment 60 of an excimer laser annealing apparatus in accordance with the present invention. Device 60 includes a quasi-molecular laser 64 that transmits a laser beam 65. The beam 65 is transmitted through a variable attenuator 66 to the beam shaping optics 68, which sends a shaped beam 69 through a rotating mirror 70 to the projection optics 72. The projection light will project the beam onto the layer 22 at a non-normal incidence angle, as previously described. The glass panel 24 containing the layer 22 is supported on a translation stage 62 that moves the layer and panel in a direction RD relative to the projected laser beam.

The device 20 described above is placed above the layer 22. The processing unit 54 determines the amplitudes of the TD ridge diffraction and the RD ridge diffracted light component observed by the detected component 52, and an electronic checklist caused by the experimental curve 譬 as shown in the curves of FIGS. 6 and 7. Whether the layer has been crystallized with a pulse above or below the OED.

Typically, the energy density (pulse energy or process ED) in the projected laser beam is initially controlled at the calibrated OED. However, the transmitted energy density may shift over time, which is typically recorded as one of the OED's significant shifts. If the OED has been shifted to a lower value than the calibrator, the ED will be lower than the OED; there will be one in the two directions mentioned above. The lower ridge density; and therefore, the amplitude of the two diffracted signals will decrease. The first signal will be sent by processing unit 54 to attenuator 66 to reduce the pulse energy delivered to the layer. If the OED appears to have shifted to a higher value than the calibrator, the ED will be lower than the OED of the instant; there will be a lower RD ridge density relative to the TD ridge described above; and, therefore, The RD ridge diffraction amplitude will decrease and the TD diffraction amplitude will remain the same. The first signal will be sent by the processing unit 54 to the attenuator 66 to appropriately increase the pulse energy delivered to the layer.

Of course, the above calibration procedure does not have to be done automatically using the feedback settings of Figure 9. Alternatively, processing unit 54 may send information regarding significant shifting of the OED to display on an operator on a monitor, and the operator can manually adjust the pulse energy delivered to the layer 22.

Figure 10 is a schematic illustration of another preferred embodiment 60A of an excimer laser annealing apparatus in accordance with the present invention. Device 60A is similar to device 60 of FIG. 9, except that the diffraction measuring device 20 is replaced by an alternative diffractive measuring device 21 that includes a directional light source 80, such as a laser beam 82. Light from the laser is incident on layer 22 at a non-normal incidence angle, as shown in FIG. 10, resulting in a reflected beam 82R and diffracted light 84. It will be a diffracted beam from the TD ridges and RD ridges as previously described with reference to apparatus 20 of Figures 8 and 9. The reflected beam 82R is selectively blocked by the stop 38 and the diffracted light is detected by the detecting element 52 and can be processed by the processing unit 54 as previously described, depending on the form of the detecting element 52.

Thus, the method and apparatus of the present invention can be used to find an OED from a panel comprising multiple scans each at a different ED, for example having an ED of 10, 5 or even ² mJ/cm ² respectively. A microscope in accordance with the present invention can be mounted in an annealing chamber of a laser annealing apparatus. The microscope can include a zoom lens assembly to vary its magnification. The panel can be scanned under the microscope to allow the panel to be measured in one or more locations under the same conditions. The microscope can additionally be provided with a platform that can cause movement in the lateral direction. An autofocus device can be attached, but not for a cone image because it has a greater depth of focus than the ELA process. Fully crystallized panels can also be measured at one or more locations (online or off-line) to detect the quality of the process so that crystallization of other panels can be aborted if desired. If sufficient measurements are completed, a map of mura (mura, the shadow of uneven brightness) can be obtained.

In the above embodiment of the invention, the OED is determined by the measured amplitude of light from a sample diffracted in one or both of the mutually perpendicular directions, such as shown in FIG. One of the potential problems with this approach is that the measured amplitude changes will occur for reasons other than ED changes. For example, degradation of the sensing component changes the measured amplitude. Also, there may be some spatial variation in crystallization on a sample that may cause spatial variability in the measured diffraction amplitude used to calibrate the same crystallized ED. The following description, with reference to Figures 11A and 11B, is an illustration of an alternative method of data acquisition and processing that can reduce, if not all, the aforementioned potential problems.

Figures 11A and 11B are micrographs of crystalline samples made with a transmissive polarization microscope. Figure 11A is made of a layer of germanium treated (irradiated) with an ED below the OED. Figure 11B is a laser beam as in Figure 11 The fragment is formed by a layer of ED that is processed above the OED. The microscope objective is a 1.25x objective. Because of this, the images of Figures 11A and 11B show a larger area than that shown in Figure 4, the images of which are taken with a 10x objective and are much larger than the images used to create Figures 6 and 7. They are taken with a 20x objective.

The horizontal dimensions of the images of Figures 11A and 11B are approximately 9 mm. With this large field, significant vignetting can occur and the brightness gradually decreases toward the edge of the image. A soft flat field correction is applied to correct it. The images will be filtered using a commercially available field map to obtain the green channel, but the blue filtered image or even the unfiltered image will have a similar appearance. Alternatively, the image can be obtained by a microscope provided with a color filter that can, for example, only filter out green or blue light. No polycondensation spacers were used for the microscope illumination. These images have high brightness and the camera shutter time is less than 100 milliseconds (ms).

The images are created in diffracted light with zero order transmitted light blocked by a polarizer and a analyzer that are rotated 90 degrees. The rotation of the polarizer and analyzer relative to the sample causes the extinction shadows to be rotated away from the diffraction belt, as previously described. This results in a microscopic image having a high density of ridges in the layer to circulate the region of light, with regions of the layer having a low density of ridges that also diffract the light. Very high contrast.

As previously mentioned, one difference in ridge density will correspond to one of the ripple densities and will correspond to one of the grain structures. Therefore, the contrast in the images indicates a heterogeneity in the layer. This heterogeneity Generally referred to as brightness unevenness (mura). The horizontal stripes in the images are in the scanning (scrolling) direction and are referred to as scanning marks. Scanning marks often occur due to slight ED variations along the long axis of the laser beam. These ED variations can cause a local offset of one of the process windows. The reason is that the scan marks below and above the OED will reverse each other. The contrast of these scan marks is found to be very stable over time and is typically only altered by realignment, contamination or clear results of the light components in the beam delivery system in an ELA system. The vertical lines in the images are in the direction of the line (the long axis of the laser beam) and are referred to as shot marks. Below OED, most of the shots are made up of wide-striped stripes, and above OED, additional, prominent stripes are typically observed.

The horizontal and vertical lines may be described as periodic for the purposes of the description and the appended claims, but this does not necessarily mean that the periodicity is regular. Moreover, the terms such as horizontal and vertical are used merely for convenience of describing lines that are perpendicular to each other, and not to indicate the actual orientation of one of the lines when implemented.

Figure 12 is a schematic illustration of how the average brightness of the inline and horizontal rows of image pixels can be used as one of the respective measured and scanned traces in an image. This image was taken in an ED range similar to the image of Figure 11B. The graphs G2 and G1 in Fig. 12 schematically show the measurement amplitudes of all the pixel in-line and all pixel cross-sections in the image, respectively. One of the individual contrast values can be determined, for example, by subtracting the lowest measured value from the respective highest measured value, or by taking the standard deviation of the amplitude variation near the average amplitude. As mentioned earlier, the image in this example was created using a soft green filter. This will enhance the contrast of the scan marks.

Figure 13 is a graph schematically showing the measurement contrast (curve A) of the scan marks obtained from, for example, the image of Figure 12 as a function of ED. This chart is obtained from an image that is filtered to show only the green channel. Also shown schematically for comparison, the green channel luminance (curve B) and the blue channel luminance (curve C) for the polarized light microscope image are similar to the curve of FIG. The magnitude is a function of ED; and one of the curves B calculates the absolute value of the slope (curve D). It can be seen that this measurement contrast (curve A) presents a very clearly defined minimum at an ED of about 15 mJ/cm ² (the OED). It can also be seen that the scan mark contrast (curve A) generally follows the form of the calculated slope (curve D), which also exhibits a clearly defined minimum at the ED as the minimum of curve A.

This is consistent with this observation, that is, the scanning trace contrast is mostly caused by the long-axis energy density variation in the linear beam. According to the relationship shown by the brightness curve B, this energy density variation is expected to cause a change in brightness. Therefore, scanning one of the measured values of the scan will produce information about the slope of the ED curve, and by this relationship, the actual processing ED will be related to the desired ED or the OED.

Fig. 14 is a graph similar to the graph of Fig. 13 showing the measurement and calculation magnitudes and the like as described above with reference to Fig. 13, but for the layer crystallized using micro smoothing. Micro smoothing is a technique used to reduce the contrast of the scan marks. The detailed description of the technology is not essential to the understanding of the principles of the invention and, therefore, is not presented herein. The smoothing technique is described in detail in the U.S. Patent No. 7,723, 169 issued to the assignee of the present application, the entire disclosure of which is hereby incorporated.

Comparing the curves A of Figures 13 and 14, it can be seen that the measured contrast is lower than in the case where micro smoothing is not used, and the minimum value of the contrast is less well defined. Nonetheless, a scan trace contrast can be measured, which faithfully increases or decreases with the slope of the luminance curve, as can be seen by comparing curves A and D of FIG. The edge is that, by this relationship, a scan mark contrast can be used as a measure of the actual processed ED relative to the desired ED or the OED for monitoring and control of the crystallization process described above.

It should be understood here that the polarized microscopic images described above for comparative measurements can be used even when one of the layers on one side of the board is being processed by ELA. When ELA, the panel moves in the scanning direction, that is, in exactly the same direction as the scanning traces. Therefore, even though the shot contrast may be subject to panel movement during a camera exposure, the scan trace contrast will not be affected.

Scanning trace contrast is very stable but very weak. Scanning trace contrast can be enhanced by introducing an optical defect, such as a thin metal wire, into the beam delivery system with care. This optical defect will cause a shadow and or a diffraction fringe in the film being processed. As long as the contrast from the optical defect is not strong, it can still be a measure of the slope of the brightness curve and can thus be used to compensate for the ED shift. The area scanned by this segment of the laser beam will have a lower consistency and may be better positioned to locate one of the panels that is not used for the active matrix of the display. For example, it can be placed at the edge of the panel or at a location that is scored and separated behind the display panel. Alternatively, the scan trace contrast will be enhanced by the temporary stop of the micro smoothing. Also, it may be better to take the picture The area of the ring is positioned outside of the area that is used to make the display.

It should be understood that although the present invention has been described with reference to the evaluation of the crystallization layer of ELA and A-ELA crystallization, the present invention is also applicable to the evaluation of the crystallization layer of other semiconductor materials. For example, a layer of germanium (Ge) or an alloy of tantalum and niobium can also be evaluated.

In general terms, the invention has been described above with reference to its preferred embodiments. However, the invention is not limited to the embodiments described and illustrated. The present invention is limited only by the scope of the appended claims.

20‧‧‧Evaluation device

22‧‧‧ Crystallized layer

24‧‧‧glass panel

26‧‧‧Microscope

28‧‧‧Light source

29‧‧‧beam

30‧‧‧Poly partition

32‧‧‧Optical components

34‧‧‧Reflected beam part

36T‧‧‧Diffraction beam part

38‧‧‧stops

39,40‧‧‧ Filter elements

50‧‧‧Detection unit

52‧‧‧Light detection components

54‧‧‧Electronic processor

Claims (7)

A method for evaluating a semiconductor layer that is at least partially crystallized by exposure to a plurality of laser radiation pulses having an energy density on the layer, the crystallization occurring on the layer in a first direction a periodic surface characteristic of the first group and a periodic characteristic of a second group in a second direction perpendicular to the first direction, the form of the periodic characteristics of the first and second groups being dependent on the semiconductor The energy density of the laser radiation pulses exposed by the layer, the method comprising: illuminating a region of the crystallized semiconductor layer; recording periodic characteristic diffraction from the illumination region by the first and second groups a microscope image of the illumination region in the light, the recorded image comprising periodic image features of the horizontal and vertical groups corresponding to periodic characteristics of the first and second groups of the illumination region of the layer; And determining the energy density from a comparison of at least one of the periodic image features of the horizontal and vertical groups. The method of claim 1, wherein the microscope image is a transmission microscope image. The method of claim 1, wherein the microscope image is formed by a plurality of horizontal rows and a plurality of in-line pixels, and the comparison is performed by measuring amplitudes of all of the horizontal and all inline of the pixels in the image. It is determined by taking the standard deviation from an average amplitude as one of the comparisons of the periodic image features. The method of claim 1, wherein the microscope image is a transmission microscope image formed by a plurality of horizontal rows and a plurality of in-line pixels, and the comparison is performed by measuring all horizontal and all in-line pixels of the pixels in the image. The amplitude of one is determined using the difference between the measured highest and lowest amplitudes as a measure of the contrast of the periodic image features. The method of claim 1, wherein the image is a green filtered image. The method of claim 1, wherein the image is a blue filtered image. The method of claim 1, wherein the one of the measured contrasts represents an optimum energy density for the crystallization.